Paul Krugman: Cognitive Capture

Krugman takes on MMT.

Let’s have a more or less concrete example. Suppose that at some future date — a date at which private demand for funds has revived, so that there are lending opportunities — the US government has committed itself to spending equal to 27 percent of GDP, while the tax laws only lead to 17 percent of GDP in revenues. And consider what happens in that case under two scenarios. In the first, investors believe that the government will eventually raise revenue and/or cut spending, and are willing to lend enough to cover the deficit. In the second, for whatever reason, investors refuse to buy US bonds.

“Investors” don’t buy bonds because they believe or do not believe in future primary surpluses. They price the bonds, and they do so based on arbitrage. If the central bank is setting a rate 3%, then bonds will be attractive to purchase if they yield a rate in excess of that, when adjusted for risk.

The second case poses no problem, say the MMTers, or at least no worse problem than the first: the US government can simply issue money, crediting it to banks, to pay its bills.

But what happens next?

We’re assuming that there are lending opportunities out there, so the banks won’t leave their newly acquired reserves sitting idle;

Here is the second problem: We are assuming that there are lending opportunities out there.  What on earth does that mean? Except for extreme circumstances — e.g. a credit crunch/increasing perceptions of risk — Banks lend as much as the market is willing to borrow at all times, at the given rate, which is controlled by the central bank. Banks post the rate, and borrowers borrow all they want at that rate. Therefore Krugman is assuming that the demand for borrowing is shifting to the right while the rate remains fixed.  Why?  And even if it does, the government can respond by increasing the lending rate.

The lending rate is not a function of the quantity of reserves, it is a function of the marginal price of reserves, but the marginal price of reserves can be set independently of the quantity of reserves. For example, Canada does this by paying one rate on reserves and charging another rate for lending reserves.  The rates it pays or charges are not constrained by the total quantity of reserves.

they’ll convert them into currency, which they lend to individuals.

No, banks do not “lend currency”. Currency is paper notes and coins. Currency is created when a household makes a withdrawal from a bank, converting its deposit into currency, in which case the bank purchases currency with its reserves and supplies it to the withdrawing households.  Again, the quantity of currency is demand determined, which makes it difficult to argue that banks or the government can force households to hold more currency than they want. The withdrawing household, unless it stores the currency in its mattress, spends it, and someone else deposits that currency in another bank, which can then purchases reserve with it. Currency is a stand-in for check-writing — it should be ignored as the relevant quantity is the sum of deposits + currency, with the exact composition a function of household preferences for paying by cash or check.

So the government indeed ends up financing itself by printing money, getting the private sector to accept pieces of green paper in return for goods and services. And I think the MMTers agree that this would lead to inflation; I’m not clear on whether they realize that a deficit financed by money issue is more inflationary than a deficit financed by bond issue.

For it is. And in my hypothetical example, it would be quite likely that the money-financed deficit would lead to hyperinflation.

Why? Notice the logical break?

Somehow (unicorns?) borrowers will borrow more even though the interest rate has not changed, merely because banks have more reserves. Then, after borrowing more, they withdraw those funds as currency (again, for some unexplained reason). Then, once they borrow large amounts and take delivery as currency, they find themselves with more cash than they want. After that, instead of rationally re-depositing the cash, they decide to rush out and buy goods with this extra cash (which they didn’t want, but nevertheless decide to withdraw). Never mind how they will pay their loans back if they use the proceeds to purchase consumption. Never mind why they would withdraw cash if they didn’t want it. Never mind why they suddenly demand to borrow more at the same lending rate.

Now it is easy to make fun such an argument — it’s a great example of cognitive capture.

Basically, deep down, Krugman doesn’t really believe in a financial sector — he doesn’t believe in inside money.

He thinks all money is outside money, created by the central bank, and that there is a demand for this outside money that, in comparison with the supply provided by the central bank, determines the price level, rather than thinking of inside money, which is demand determined at the going rate. The real issue is the rate of interest, not the quantity of reserves. The real issue is  private sector’s borrowing costs, not the how the government finances itself.

The non-financial sector doesn’t care how many reserves there are. It cares about borrowing rates and deposit rates.  Neither the banks, nor government can force households to hold more deposits or currency than they want under a modern credit-based financial system. Rather, the government can force banks to hold whatever level of reserves it requires, and simultaneously the government can set any marginal price for those reserves.  There is no reason why the marginal cost of reserves needs to fall if the quantity of reserves increases, as both the quantity of reserves and their marginal price are orthogonal policy variables.   Typically, central banks set the marginal price and have the quantity of reserves be as small as possible. Some nations (e.g. Canada) set the quantity of reserves to be zero (actually, a few million) and only control the price. QE is an example of controlling both the quantity and price. You can have, btw, QE with a high and non-zero marginal price of reserves, simply by either paying banks a high level of interest on reserves, or imposing asset taxes on banks that force them to not lend money out unless a certain (pre-tax) return can be earned — enough of a return to pay the tax.


Paul Krugman: Cognitive Capture

12 thoughts on “Paul Krugman: Cognitive Capture

  1. mpr says:

    Surely the lending rate isn’t just a function of the marginal price of reserves at that point in time. Long tsy rates effect this too.

    So how about the following for an explication of what PK is trying to get at: The govt. spends, but issues no new bonds. The spending becomes excess reserves, which are used to buy tsys. This causes the price of tsys to rise (since there were no new ones created) and yields to fall.

    This causes bank lending rates to fall, increasing loan demand, and hence increasing AD in excess of the initial spending.

    All else being equal, loan demand continues to be elevated until the level of required reserves has increased by the amount of spending in question. (So that the price of treasuries has fallen back to their original level at this point). However, if this process has created increased inflation then inflationary expectations could lead to higher loan demand even past this point, so tsy yields rise further etc. and you get your hyperinflationary spiral.

    1. No, the lending rate depends on many factors, such as risk, bank regulation, expected future short rates, etc. But you are not taking your argument all the way through. Reserves are not a consumption good. They do not become unavailable after they are spent — they remain in the system. So if all the CB did was OMO, then as soon as even a small amount of excess reserves were created — say, 1/100 of 1% of the base — then short term rates would fall and keep falling to zero until the CB did some counter OMO to get rid of the excess reserves. In such a system is it impossible for the government to both print money and control interest rates.

      But we do not live in such a system – we can pay interest on reserves. If reserves pay an interest of 3%, then banks, even with excess reserves, would only bid down short bonds to 3%. If that is the target, then we can maintain a 3% target rate independently of the quantity of reserves.

      And we don’t need to even do that. It’s enough to tax banks — say we tax them 3% on any assets that they hold (other than reserves). Now, banks would never lend to anyone at a rate less than 3%, nor would they bid down bond yields below 3%, regardless of the quantity of reserves.

      As you can tell, the response of the system to an increase in reserves is *heavily* dependent on the institutional infrastructure and regulation in place. There is no god given relationship between the quantity of reserves and interest rates — it is a policy variable. You can’t approach this problem with a simple model in which there are no banks or no deposits and hope to describe what would happen.

  2. mpr says:

    “then short term rates would fall and keep falling to zero until the CB did some counter OMO to get rid of the excess reserves.”
    Ah, so the banks cant get rid of the excess reserves themselves by trading tsys with each other ? Ok, but this makes the argument even more direct: To do OMO the CB needs to swap the reserves for tsys, of which it will eventually run out. Otherwise the FFR falls to 0, which reduces lending costs etc.

    “But we do not live in such a system – we can pay interest on reserves.”
    This is like funding with overnight T-bills. Is it really reasonable to completely fund this way ? Maybe. I wonder if there would be any side effects. Why do they choose not to do it this way now ?

    “And we don’t need to even do that. It’s enough to tax banks”
    I like this, but its a major change. I wouldn’t say PK is ‘wrong’ because he didn’t consider this, although maybe its worth pointing out as a policy option.

    I think MMT would have been better served if people didn’t focus so much on the fact that what PK is saying doesn’t literally correspond to the mechanics of the banking system. The cause and effects he is talking about do have some basis in reality, even if he is not describing them in the MMT paradigm.

    Its possible to make all the substantive points you made
    in his ‘money multiplier’ paradigm – pay IOR, tax the banks.

    Your blog is great BTW.

  3. ” To do OMO the CB needs to swap the reserves for tsys, of which it will eventually run out. Otherwise the FFR falls to 0, which reduces lending costs etc”

    Yes! In a system in which the CB manages the overnight rate only via OMO, the government cannot monetize debt unless rates are at zero.

    But this is not an immutable law of money demand, it is the direct outcome of a specific institutional arrangement. You can change the arrangement — e.g. to paying interest on reserves.

    “Maybe. I wonder if there would be any side effects. Why do they choose not to do it this way now ?”

    They do, but only recently. Canada, Australia, NZ use a corridor system rather than OMO. I don’t think the underlying arrangement is the explanation of why we do not monetize more, rather we choose not to, and then build a system compatible with that belief.

    But the point I was making is that there is no stark rubicon from monetizing debt due to pure economic reasons. At the end of the day, there are limits to seignorage, but crossing those limits does not result in some hyperinflationary explosion — households are already willing to hold the entire sum of the federal debt held by the public as zero maturity deposits.

    For some reason, we allow the banks to earn this seignorage income and create as many deposits as they want, while we worry about hyperinflation if the government creates more money.

    1. Thanks, Scott. I don’t think there is anything new here, but I am surprised that Krugman really is not thinking in terms of banks at all. Both he and Nick R. think of MZM as something controlled by the central bank, or they just don’t have models in which there is an MZM at all. Perhaps I am wrong, and they do have such models, but I can’t detect it in the public discourse.

  4. J.V. Dubois says:

    First, sorry for late response to this topic. I discovered your blog only recently, following from Nick’s and Andy’s blogs. I assume that you are proponent of MMT, which I briefly studied for last months. I have several questions for which I did not received satisfactory answers on Bill Mitchell’s blog – as for some reason Bill and people there are very hostile to anybody asking genuine questions, trying to mire such questions down into tautology and word games. They lack the openess Nick (and I believe you) shows in his blog – to anticipate the question asked and answer it in a way understandable to the party asking.

    So to my questions. What is your position on this post from Scott Sumner discussing the price level? Do you agree that the price level is tied to base money, not to bonds?

    The other thing I hear a lot from MMTers is that governments “needs not to be financed” meaning that they can print whatever money the want, without issuing bonds. However MMT clearly acknowledges that inflation can become a problem if government spending by printing money reaches certain level. When in one discussion I used term of financing government expenditures by *future* taxes and/or inflation, Bill get on me that this is not how “financing” is defined in English. But clearly, if Scott is right that price level is affected by Base Money (not by bonds), and if government spends by directly printing money, such operation is inflationary.

    When presented with this idea, MMTers say that government may come with a solution – start offering interest on reserves. So you will be able to sterilize part of the monetary base by using it as store of value. But to me this is only a trick. If government continues printing money to spend, it will eventually have to offer higher and higher interest on reserves to prevent inflation. How is this effectively different from issuing bonds?

    The third question is of magnitude. Government is capable of gathering taxes in a range of 0-100% GDP. If taxes should be solely used for managing aggregate demand how would you be able to respond to large shocks? Clearly, changing tax structure by 10s of percentage points is not an easy feat.

    1. mpr says:

      Sumner’s example with the ship dropping off the money is one of my favorites. Its very funny because he doesn’t seem to realize at all that the answer is “they burn the money to keep warm” or perhaps “they use the money as toilet paper”.

  5. JV,

    “Clearly, changing tax structure by 10s of percentage points is not an easy feat.”

    It is if you create the system anticipating that.

    There is no fundamental difference between monetary adjustments and fiscal adjustments. They all attempt to do the same thing – stop people spending when there is too much demand and get people spending when there isn’t enough demand while at the same time ensuring the supply system can feed the current demand.

    Interest on Reserves is almost exactly the same as Bonds, but saves all that messing around with refinancing. In fact one if you look up Monetary Sovereignty you’ll find that branch of Post Keynesian economics advocates using Interest rates as the control mechanism exclusively.

    MMT rejects that because it still causes a build up of money.

    The feedback mechanism is simple in MMT. You look at the unemployment rate and the demand component of inflation and you adjust taxation and fiscal policy in some manner until unemployment is eliminated and the demand component of inflation remains benign. The floating exchange rate then provides the degree of freedom necessary for you to hold the other two variables stable.

    Once you’ve got to that then you operate a counter cyclical fiscal structure that expands government money as private money contracts, and contracts government money as private money expands.

    Add to that tight capital controls on the banks so they can’t lend too much and you have a credit cyclical system that can be controlled.

    In terms of system design, MMT advocated using buffer stocks to absorb shocks. That’s what the Job Guarantee is, and a similar proposal has been suggested for needed commodities to handle supply shocks.

    Mostly its about beefing up the automatic stabilisers so they have the strength to return private sector spending to its ‘ignition’ point without further intervention.

    In terms of emergency taxation measures, it would be fairly simple to delegate authority to the central bank to raise a land levy when it feels it needs to to control inflation. That allows the central bank to drain funds from the system.

    If that gets charged to the freeholder, or the entity with first charge over the freehold, then the cost gets passed down to householders just like a rise/fall in variable mortgage rates does now.

    Interest rates can then be set low and stable at whatever level encourages the greatest amount of productive private sector investment. The market then sets the price of money within the capital constraints of the banks.

    But most importantly the amount of government deficit at any point in time will only be what is required to offset the net financial saving of the non-government sector.

    Government will back off and increase taxation/cut spending as the private sector takes over. In any decent design that will be via the automatic response of the stabilisers so that you don’t have the discretionary spending problem with politicians.

  6. J.V. Dubois says:

    Neil Wilson: Nice to discuss with you again, I remember having some discussion over at Billy blogs. If I remember it correctly, MMT (or at least Bill Mitchell) states on several ocasions that setting the interest rate is wrong. I will offer direct quote from this article of his:

    “First, monetary policy is a dubious tool to use to counter-stabilise aggregate demand. It is not entirely clear (or predictable) which way the interest rate effects will go with respect to spending. The distributional complexities of an interest rate cut (creditors lose, debtors gain) make it hard to know what will happen. Further, the policy tool is blunt, indirect, cannot be targetted and is subject to unknown lags).”

    So you see, it is very strange for the same MMTers people to argue that we do not need bonds because we can pay interest on reserves – when they reject the very idea of using interest rate at all. Actually I remember several Bill’s blogs where he advocated solely inflation “financed” government expenditures. Given that there are virtually no countries with government share less than 20% of the GDP, we would expect the inflation in double digit territory.

    And now the weird part – as you know, the MMT tells us that the interest rate should be zero. Why? Because if government does not sterilize the money (via bonds or interest on reserves) than according to the theory banks would have systematic overhang of reserve. No matter what they do with it (buying private bonds, stocks etc.) they just cannot get rid of it collectively. MMTers claim that this means that the interest rate drops to zero. Private or “vertical” money has no channel for escape. Even if there is 1 dollar left in the system, this will be lent for zero interest on margin. So you have combination of 0% interest rate on margin and double digit expansion of the monetary base via direct government purchases. What would that leave us with? Only massive inflation.

    So I will dedicate this last paragraph to MMT contradictions (at least as I see them after reading Billy Blog).

    1. They reject the existence of Long Term Budget Constraint because government may print and spend whatever money they want. If you object that this would lead to high inflation, they will point out that you may use taxes to balance it. But then LTBC is valid, isn’t it?

    2. No MMTers say, because you may also start paying interest on reserves to prevent inflation. But then standard monetary policy – which manipulates AD via interest rate is valid, isn’t it?

    3. No say MMTers because monetary policy is a dubious tool with no way ensure if it stimulates AD or not. It has to be government which does the job of AD stabilization. Why, you ask.

    1. Because government does not face LTBC and it may print whatever money it wants … and the circle starst again.

    I always get into some verions of this circle. It is like Whac-A-Mole game. I still wait to see if any MMTer can explain how these things relate in a comprehensive way.

  7. You get into some version of the circle because you haven’t fully understood the system.

    If you continue to view it from a standard viewpoint you won’t get it. You have to move position.

    The problem that underlies the LTBC idea is handled simply in MMT by focussing on today’s inflation and today’s unemployment rates. MMT has never said you can print and spend what you want. You can only buy what is for sale today that isn’t required by anybody else. If you consider how a Job Guarantee system works you’ll see how the spending backs off as the private sector recovers – automatically and counter-cyclically.

    It’s essentially buffer stocks, enhanced automatic stabilisers and a focus on optimising the Real situation rather than hyperventilating about meaningless financial ratios.

    MMT’s position on the interest rate is that of Functional Finance – that it should be set at a level that ensures the optimal amount of productive investment. With the right tax regime and capital control systems on the banks it’s likely that will be at or close to zero. It depends on the nature of the country and its real production system.

    There are various political flavours of the basic idea, but in essence they are based on the same underlying description of the nominal system.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s